EPA-AA-HDV 76-2
Technical Support Report for Regulatory Action
Heavy Duty Truck Evaporative
Emissions Regulations Development
A Progress Report
John Corcoran
July 1976
Notice
Technical support reports for regulatory action do not necessarily
represent the final EPA decision on regulatory issues. They are intended
to present a technical analysis of an issue and recommendations resulting
from the assumptions and constraints of that analysis. Agency policy
considerations or data received subsequent to the date of release of this
report may alter the conclusions reached. Readers are cautioned to seek
the latest analysis from EPA before using the information contained
herein.
Standards Development and Support Branch
Emission Control Technology Division
Office of Mobile Source Air Pollution Control
Office of Air and Waste Management
U.S. Environmental Protection Agency
-------
Background
Heavy duty gasoline fueled vehicles are not currently covered by
any evaporative hydrocarbon standard, except in California where they
must meet the light duty standard of 2 grams/test. California certifies
the system by design evaluation of-the control system rather than re-
quiring confirmatory testing of completed vehicles, so information on
the performance of these systems is nonexistent.
The primary reason for the lack of a regulatory proposal to control
evaporative emissions is the lack of a comprehensive test procedure..
The adoption of an enclosure (SHED) procedure is expected to alleviate
this constraint, but requires much development before it can be applied
to heavy duty vehicles. Another important problem is the method of
insuring compliance with evaporative standards. Testing requires a
total vehicle while EPA currently certifies only engines for heavy duty
vehicles and would have to begin dealing with total vehicle manufacturers
in a complete vehicle compliance program. California has sidestepped
this problem by not requiring a demonstration of compliance. Engineering
judgment is substituted. If EPA were to merely adopt the California
program the inability to insure compliance by certification testing of
all.'possible vehicle configurations could further erode the estimates of
potential emission reductions obtainable.
Purpose
This is a status report on the Heavy Duty Truck Evaporative Emis-
sions Project. The specific goals of this project are as follows:
1. Develop a preliminary test procedure using the large enclosure
(SHED) at MVEL.
2. Use the preliminary test procedure to get initial estimates of
the magnitude of the evaporative losses from uncontrolled and controlled
heavy duty trucks.
3. Apply these initial test results to a revised air quality
impact assessment.
4. Estimate the performance potential of current control systems
through an in-house technology assessment program.
5. Evaluate the alternative compliance assurance strategies.
6. Develop a complete regulations package including standards,
and a finished test procedure.
-------
-2-
The areas that will be covered in this report are:
1. Initial test procedures.
2. Initial test results.
3. Procedure development.
Summary and Conclusions .
When tested in a manner similar to the proposed 1978 LDV SHED test
for evaporative emissions, the uncontrolled heavy duty gasoline fueled
trucks tested emitted an average of 44.63 grams/test HC which is equiv-
alent to 4.04 grams/mile HC. Trucks with control systems for the
current California standard show a significantly lower evaporative loss,
14.78 grams/test or 2.00 grams/mile, however they are still high when
compared with the LDV exhaust and LDV-LDT evaporative emissions standards.
The one Diesel fueled truck tested indicates that evaporative
losses from Diesel fueled trucks are low enough/(1.65 grams/test,
0.9 ,grams/mile) that they are not a concern at "this time.
The one hour diurnal loss test does not reasonably represent an
actual diurnal cycle, therefore the measured losses may be less than
those occurring from in-use vehicles. It appears that lengthening the
diurnal loss test, to approximately 3 hours, may solve this problem.
This will yield significantly higher losses than reported here, indicating
that the HDV evaporative losses may be an even greater problem than
anticipated.
Discussion
Initial Test Procedures
MVEL has an evaporative emissions enclosure (SHED) that will accom-
modate vehicles up to 10 feet wide, 11 feet high, and 33 feet long. It
was decided to use this SHED for all heavy duty evaporative emissions
testing. The SHED is similar to the one used for light duty vehicle
testing. Hydrocarbon concentrations are measured with a Flame lonization
Detector (FID).
SAE Recommended Practice J171a was chosen as an initial test proce-
dure, however, the following changes were made:
1. The vehicle background emissions were not separated from the
total evaporative emissions. Background emissions are a source of hydro-
carbons with an effect on air quality and should be considered part of
the evaporative losses.
-------
-3-
2. In place of a chassis dynamometer driving cycle between the
diurnal and hot soak loss tests, the trucks were driven over an 8 mile
road route. The route, shown in the appendix, was intended to provide a
consistent means of bringing the vehicles up to operating temperatures
for the hot soak test.
Aside from these changes, the test procedure is the same as SAE
J171a. There is a diurnal test where the liquid fuel is heated from
60°F to 84°F in 60 minutes and a 60 minute hot soak test immediately
following a driving cycle.
Initial Test Results
Nine gasoline fueled trucks, with Gross Vehicle Weight Ratings in
excess of 6,000 Ibs., without evaporative emission control systems were
tested by the above procedure. The results vary from 21.43 to 75.8
grams/test and are detailed in Table 1. The equivalent grams/mile HC
was determined from the assumption that heavy duty vehicles experience 1
diurnal cycle and 9 hot soak cycles (trips) per day, and travel 36.2
miles. This duty cycle is based on studies of in-use vehicles in New
York City. (EPA Contract Number: 68-01-0414)
-',This data shows several points worthy of discussion;
First - The trucks with two 50 gallon tanks had approximately twice
the diurnal losses of the truck with one 50 gallon tank. Since the
tanks were approximately the same shape this was expected, however,
for the trucks with single fuel tanks, the losses "are not directly
proportional to the volume of the tank. The losses from the 36
gallon tank averaged 21.33 grams, while the losses from the 75
gallon tank averaged 27.01 grams. Therefore, there must be other
factors, besides tank volume, contributing to the diurnal losses.
Second - There seems to be no pattern to the hot soak losses.
These losses are affected most by carburetor design and underhood
temperatures (which are a function of engine compartment packaging).
It is unlikely that a pattern would show up in such a varied sample
of trucks.
Third - The hot soak loss is the dominant term in the determination
of the equivalent grams/mile HC. Comparing the Chevrolet pick-up
truck and van the total losses for the pick-up are less than 10%
greater than those of the van. However, the equivalent grams/mile
HC for the pick-up is 50% higher than the van's. This is due to
the higher hot soak losses of the pick-up and the large number of
trips (hot soaks) used in calculating the equivalent HC.
\
Four gasoline fueled trucks with evaporative emissions control
systems designed for use in California were also tested. The results of
-------
-4-
Table 1
. EVAPORATIVE EMISSIONS FROM
•!.Truck Fuel Engine
(GVW) Capacity (CID)
UNCONTROLLED
Diurnal
GASOLINE FUELED TRUCKS
Loss (g)
Hot Soak
Total
Equiv. HC
Cms /mi .
Ford N-600 50 Gal.
(EPA-162).
(22,000)
Ford 25 Gal.
L-800
(31,000)
Ford 75 Gal.
LT-900
(46,000)
"'*
Ford Van 22 Gal.
E-150
(6,300)
Chevy Pickup 20 Gal.
C-20
(8,200)
Chevy Van 36 Gal.
(7,900)
CMC - 6500 100 Gal.
6,500 (2 X 50)
(27,500)
CMC - 6500 100 Gal.
Sierra (2 X 50)
(23,000)
CMC - 6000 100 Gal.
(24,000) ' (2 X 50)
361
V-8
Avg. =
391
V-8
Avg. =
534
V-8
Avg. =
300
1-6
Avg. =
292
1-6
Avg. =
Avg. =
427
V-8
Avg. =
366
V-8
Avg. =
350
V-8
Avg. =
• 25.86
21.90
19.69
22.18
22.41
17.63
17.41
16.48
18.13
17.52
17.43
28.42
24.62
26.86
28.12
27.01
18.19
19.68
18.94
18.13
18.44
18.29
22.36
20.29
21.33
52.44
52.74
51.87
52.35
48.83
49.99
52.06
53.10
53.18
52.56
51.62
57.88
57.88
13.41
16.21
8.41
7.93
11.49
4.02
3.67
4.91
2.18
5.19
3.99
12.76
14.72
13.74
9.15
11.53
10.34
10.44
10.03
10.24
4.79
5.56
5.18
24.02
23.87
22.46
23.45
22.50
15.75
17.99
17.01
20.44
18.74
17.77
16.67
17.22
39.27
38.11
28.10
30.11
33.90
21.65.
21.08
21.39
20.31
22.71
21.43
41.18
39.34
40.75
27.34
31.21
29.28
28.57
28.47
28.52
27.15
25.85
26.50
76.46
76.61
74.32
75.80
71.50
67.81
71.09
70.19
73.00
70.36
75.65
75.10
3.48
1.47
4.16
. 3.09
3.05
1.88
7.28
6.09
5.88
Group
Average = 4.04
-------
—5—
[Table 2 -
EVAPORATIVE EMISSIONS FROM CONTROLLED
Truck Fuel Engine
(GVW) Capacity (C.I.D.)
Diurnal
GASOLINE FUELED TRUCKS
Loss (g) Equiv. HC
Hot Soak
Total Cms /mi.
Ford
LN-700
(25,500)
Ford
F-250
(6,900)
Ford
L-900
(56,000)
I.H.C.
Fleetstar
2010A
(34,500)
50 Gal. 361
V-8
Avg. =
41.5 Gal. 390
(22.5 4- V-8
19) . Avg. =
50 Gal. 534
V-8
Avg. =
126 Gal. 537
(2 X 63) V-8
Avg. =
12.49
12.64
13.63
12.92 .
3.87
3.64
3.76
4.08
4.48
5.28
4.61
2.78
14.09
13.47
9.37
5.26
8.99
5.78
4.00
4.08
4.62
11.93
13.51
12.72
6.63
3.97
3.72
4.77
5.04
5.94
6.07
9.05
9.25
4.68
6.67
18.27
16.64 1.51
17.71
17.54
15.80
17.15 3.27
16.48
10.71
8.45 ' 1.31
9.00
9.38 .
7.82
20.16
22.52 1.91
18.62
9.95
15.7
Group
Average =2.00
-------
-6-
these tests varied from 9.38 to 17.54 grams/test and are shown in Table
2. For the three Ford trucks, the test results are quite consistent.
However, the diurnal losses from the I.H.C. Fleetstar varied widely and
no explanation has yet been found.
In order to make a rough comparison between the controlled and
uncontrolled trucks, the mean of the total loss and the mean of the
equivalent HC are shown in Table 3.
Table 3
Comparison of Uncontrolled and Controlled Trucks
Mean Values
Total Loss (Cms/Test) Equivalent HC (Cms/Mi)
9 Uncontrolled
Trucks 44.63 4.04
4 Controlled
Trucks 14.78 2.00
"'There is a significant reduction in hydrocarbons with the evaporative
control systems. However, the magnitude .of the losses from the controlled
trucks is still quite high.
These initial tests indicate that an improvement of approximately
2 grams/mile HC is possible with the control systems- designed for the
current California regulations. However, these tests also indicate that
even with the current control systems heavy duty evaporative emissions
will continue to have a significant impact on total HC emissions. At
14.78 grams/test or 2 grams/mile individual gasoline-fueled controlled
vehicles have much higher losses than emissions which will be allowed
from future light duty vehicles (See Table 4).
Table 4
Future Light Duty Vehicle Emission Standards
Total SHED Evaporative Equivalent HC
Loss (gm/test) (gms/mi)
LDV (1978 Statutory - 0.41
exhaust standard)
LDV (1979 Proposed 2.0 0.2
evaporative standard)
-------
-7-
One Diesel-fueled truck has been tested as shown in Table 5. If
this truck is representative of all Diesel trucks, Diesel evaporative
emissions regulation is not critical at this time.
Table 5
Evaporative Emissions From A Diesel Fueled Truck
Truck Fuel Loss(g) Equivalent
(GVW) Capacity Engine Diurnal Hot Soak Total gms/mi
CMC ASTRO-95 100 Gal. 8V71N 1.44 .21 1.65 . .09
(120,000 GCW) (2 x 50)
Test Procedure Development
Before testing began, it was assumed that hot soak losses from
uncontrolled heavy duty vehicles would be approximately equal to those
from uncontrolled light duty vehicles. Hot soak losses are primarily
carburetor losses and essentially the same carburetors are used on light
and heavy duty vehicles. The Los Angeles FY 71 test program showed that
uncontrolled light duty vehicles had- a mean hot soak loss of 14.86
grams/test. The nine uncontrolled trucks in Table 1 had a mean hot soak
loss of 12.71 grams/test. Therefore, the assumption is accurate.
It was also assumed that diurnal losses from uncontrolled heavy
duty vehicles would be several times greater than those from uncontrolled
light duty vehicles. This was based on the fact that heavy duty vehicles
have larger fuel tanks than light duty vehicles, and many heavy duty
vehicles have multiple fuel tanks. The Los Angeles FY 71 test program
revealed a mean diurnal loss of 25.73 grams/test for uncontrolled light
duty vehicles. The three trucks in Table 1 with single fuel tanks of
25 gallon capacity or less had a mean diurnal loss of 18.22 grams/test
and a mean fuel tank capacity of 22.33 gallons. The three trucks equipped
with two 50 gallon tanks had a mean diurnal loss of 53.95 grams/test or
26.98 grams/tank/test. There is some increase due to tank size but, the
assumption that the diurnal losses are directly proportional to .fuel
tank volume is incorrect. There must be some other variables that also
contribute to the diurnal loss.
In an SAE paper entitled "Factors Influencing Vehicle Evaporative
Emissions", paper number 670126, D.T. Wade of Esso (Exxon) Research and
Engineering Co. presents the mechanisms controlling diurnal losses and a
mathematical relationship for predicting the losses. Wade contends that,
when there is equilibrium between the liquid and vapor in the tank, the
diurnal losses are a function of the properties of the fuel, the atmospheric
pressure, the pressure inside the tank, the initial and final fuel and
vapor temperatures, and the vapor space in the tank. The test procedure
-------
-8-
keeps all of the factors constant except the vapor space, which is a
function of the tank design. Therefore, the diurnal loss should be
directly proportional to the vapor space. The test results show that
this was not so. As the vapor space increased, the measured losses also
increased but not in the same proportion as the vapor space. Wade
cautions that heating rates faster than the rate of equilibration, will
produce measured losses that are smaller than the simulation predicts.
Further investigation revealed that the one hour diurnal test cycle is
not an equilibrium process. Figure 1 shows the fuel and vapor tempera-
tures (measured at the centroids) and losses plotted against, time for an
uncontrolled 50 gallon tank. The fuel and vapor temperatures are not
equal at the start of the test and the fuel temperature rise is faster
than the vapor temperature rise. Increasing the length of the diurnal
cycle (in effect, slowing the heating rate) to 3 hours does not signifi-
cantly alter the relationship shown in Figure 1. However, it does cause
a significant increase in the measured losses, to 31.63 grams for the
tank tested.
Apparently, when the diurnal loss test is not an equilibrium process
predicting the results is currently not possible. Added to the variables
already known are the proportions of the tank (which will affect the
initial vapor temperature, the rate of increase of the vapor temperature,
and the temperature gradients in the fuel and vapor), the length of the
test (which possibly affects the temperature gradients in the fuel and
vapor, and the HC concentration of the expelled vapor), and the heating
method (which affects the temperature relationships).
It is necessary to know whether the diurnal temperature cycling of
in-use vehicles is an equilibrium process. If it is, then the current
one hour diurnal test cycle will never yield losses as great as those
from vehicles in the real world, and the test will need to be improved.
The only available data on this comes from the CRC/CAPE-5-68 program run
by Scott Research Laboratories and is shown in Figure 2. Twenty-nine
vehicles were parked outside from 8 A.M. until 12 noon and the tempera-
tures were recorded. Since the test was started at eight and ended at
noon the ambient temperature rise, 12.9°F, was only about half of the
24°F rise on which the SAE J171a procedure is based. Figure 2 shows
that in-use diurnal cycles are not equilibrium processes. If the results
of Figure 2 are extended to a 24°F ambient temperature rise then a fuel
temperature rise of 12.84°F and a vapor temperature rise of 20.09°F
would be expected. The one-hour diurnal test (SAE J171a) shown in
Figure 1 shows a 24°F fuel temperature rise and a 10°F vapor temperature
rise. Therefore, the SAEJ171a procedure does not reproduce actual
conditions.
Figure 2 also indicates that the ambient air temperature is the
forcing function for the diurnal cycle. The fuel and vapor temperature
cycles follow the ambient air temperature cycle, but there is a time lag
caused by the "thermal inertia" of the fuel or vapor. An attempt was
made to duplicate the real world conditions by letting the circulating
air in the SHED heat the fuel and the vapor. For this test it was
-------
-9-
Figure 1 - 1 Hour Diurnal Loss Test (SAE J171a)
84
CO
0]
w
CO
30
20
10
80 -
76 -
72 -
70.7-
o- 68
Q)
H
64 -
60
Fuel temp.
£ .
fuel
= 24°F
AT = 10°F
vapor
Vapor temp.
Average of
3 tests
0 10
20 30 40
Time (min.)
50 60
Figure 2 - Actual Vehicle Fuel System Temperatures
96-
92-
I"
cu
H
88-
84-
80-
76-
72-
68
AT
vapor
fuel
I |
2 ' 3 4
Time (hr.)
Source:
CAPE
5-68
29 vehi-
cle
average
-------
-10-
desirable that the fuel and vapor temperatures be approximately equal at
the start of the test. The test was performed on a 50 gallon uncontrolled
tank which was placed outside overnight to cool it below 60°F. It was
brought into the SHED and filled to 40% of rated capacity with 60°F
gasoline. The SHED was sealed, and left for 7 hours (Figure 3). The
temperature profiles do not resemble Figure 2. The main reason for this
is the constant 80°F ambient temperature. Since the vapor has a lower
"thermal inertia" than the fuel it is heated more rapidly. However, the
temperature changes are quite similar to the real world conditions,
15.8°F vs. an expected 12.84°F for the fuel and 17.8°F vs. an expected
20.09°F for the vapor. The losses for this test were 30.28 grams. The
same tank tested by. the 1 hour procedure (SAE J171a) had losses that •
averaged 20.85 grams, with fuel and vapor temperature changes of 24°F
and 10°F respectively.
It was not possible to lower the ambient temperature in the SHED so
that the vapor temperature would rise at the same rate as the fuel
temperature. So, for a subsequent test, the rate of increase of the
fuel temperature was increased by heating the fuel with a heat blanket.
The fuel and vapor temperatures were both 60°F at the start of the test,
and the fuel was heated to 84°F in one hour (Figure 4). A comparison of
the three types of tests is given in Table 7.
~ 4
Table 7
Comparison of Diurnal Test Cycles
Fuel Temp (°F) Vapor Temp (°F) Losses
Test Type Initial Final A Initial Final A (gr HC)
SAE Jl7la (Fig. 1) 60 84 24 70.7 80.7 10 20.85
SAE J171a (Fig. 4) 60 83.7 23.7 60 79.9 19.9 23.29
T = T
(fuel). (vapor).
7 Hour Soak (Fig. 3) 61 76.8 15.8 60 77.8 17.8 30.28
The variables in this series of tests were the initial and final
fuel and vapor temperatures and the test length. The two one hour tests
had essentially the same initial and final fuel temperatures, so the
only difference in the tests are the vapor temperatures. However, the
final vapor temperatures are almost equal so that the only difference is the
initial vapor temperature and, therefore, the vapor temperature change.
Doubling the vapor temperature change, by lowering the initial vapor
temperature, increased the losses approximately 12%, from 20.85 grams
to 23.29 grams. The 7 hour soak test had fuel and vapor temperature
changes that were less than the one hour test shown in Figure 4, yet the
-------
-11-
Figure 3-7 Hour Soak Diurnal Cycle
30
05
2 20
t>o
8 10
0 _
84 _
AT,, . = 15.8°F
fuel
AT = 17.8°F
vapor
Loss = 30.28 gr.
10
Time (hr.)
Figure 4-1 Hour Diurnal Test (SAP.-J17.1a) with T/f. ,.= T. ,
—& , (fuel). (vapor).
30-
x-^
CO
CO
M 20^
CO
CO
o
OJ
84 _
80
76
e 72
68'
64
60.
C .
fuel
= 23.7°F
AT = 19.9°F
vapor
Loss = 23.29 gr.
Average of
3 tests
Loss
i i i i i i
10 20 30 40 50 60
Time (min.)
-------
-12-
losses were 30% higher, 30.28 grams vs 23.29 grams. Therefore, the
length of the test is an important variable which cannot be overlooked.
This was not mentioned in Wades simulation which only applies to equilibrium
processes where the test length is always the same. Since actual diurnal
cycles and any laboratory simulations are not equilibrium processes the
length of the test is a variable that must be taken into account.
Of all the lab tests run, the 7 hour soak (Figure 3) is the best
simulation of the actual diurnal cycle (Figure 2) when the major variables
are compared (Table 8).
Table 8
7 Hour Actual Diurnal
Soak Cycle*
Fuel Temperature
Increase 15.8°F 12.89°F
Vapor Temperature
Increase 17.8°F 20.09°F
Time 7 hr >4 hr
*Extrapolating Figure 2 to a 24 °F ambient temperature rise.
Getting a better simulation of the actual diurnal cycle requires controlling
the temperature of the air in the SHED which is not possible at this
time. However, a 7 hour test is burdensome for a certification procedure,
and controlling the SHED air temperature is complex and expensive. What
is needed is a shorter test that yields representative losses.
It was mentioned earlier that lengthening the diurnal test, while
using the same 24°F fuel temperature rise, increased the measured losses
without any other change in the process. Therefore, it is possible to
lengthen the SAE J171a test until the losses equal the losses from an
actual diurnal cycle. It was not possible to measure the losses from an
actual diurnal cycle, however, for the truck tested, the 7 hour soak
yields losses similar to the real world cycle. The losses from the 7
hour soak were 30.28 grams, the 3 hour version of the SAE J171a gave
losses of 31.63 grams. A considerable amount of time can be saved,
without changing the losses, by using a longer version of the SAE J171a
test instead of a 7 hour soak. Further testing will be needed to determine
the proper length for the diurnal test.
-------
-13-
APPENDIX (Ann Arbor, Michigan)
Evaporative Emissions Test Road Route
- \1
f / f^ Mf
-------
Abstract
Initial evaporative emissions, testing of heavy duty vehicles by the
SHED technique is described and test results reported. A potential
weakness in the diurnal loss measurement method is discussed and further
testing recommended.
Approve^- Pro'ject Manager,
Heavy Duty Vehicles
.. **>&. C.CflW n^ J ' fJ-*-
Approved - Branch Chief, SDSB Approved - Division Director, ECTD
Distribution:
D. Alexander
W. Clemmens
T. Darlington
J. DeKany
C. France
C. Gray
L. Higdon
R. Kruse
M. Leiferman
J. Marzen
T. Rarick
G. Rossow
R. Stahman
E. Stork /
M. Williams J
G. Wilson
------- |